Automated brain volumetrics in multiple sclerosis: a step closer to clinical application
C Wang1, H N Beadnall2, S N Hatton1
1Sydney Neuroimaging Analysis Centre, Sydney, New South Wales, Australia Brain and Mind Centre, The University of Sydney, Sydney, New South Wales, Australia.
Journal of Neurology, Neurosurgery, and Psychiatry
|April 14, 2016
Summary
Whole brain volume (WBV) measurements in multiple sclerosis (MS) patients are crucial. NeuroQuant and MSmetrix offer comparable accuracy to SIENAX for assessing cross-sectional WBV in MS.
Area of Science:
- Neuroimaging
- Neurology
- Medical Statistics
Background:
- Whole brain volume (WBV) is a key biomarker in multiple sclerosis (MS), correlating strongly with clinical disability.
- Traditional lesion-based metrics in MS are less robust than WBV.
- Several algorithms exist for WBV measurement, necessitating comparative analysis.
Purpose of the Study:
- To compare the accuracy and precision of NeuroQuant and MSmetrix for cross-sectional WBV assessment in MS patients.
- To evaluate the statistical agreement between SIENAX, NeuroQuant, and MSmetrix for WBV measurements.
Main Methods:
- Analyzed MRIs from 61 relapsing-remitting MS patients and 2 clinically isolated syndrome patients.
- Calculated WBV using SIENAX, NeuroQuant, and MSmetrix.
- Assessed statistical agreement using linear regression and Bland-Altman plots, calculating precision and accuracy.
Main Results:
- NeuroQuant and MSmetrix demonstrated high precision (r=0.983 and 0.992) versus SIENAX.
- Accuracy (Cb) was 0.871 for NeuroQuant and 0.994 for MSmetrix against SIENAX.
- Volume differences were 5.5% (NeuroQuant) and 1.0% (MSmetrix) compared to SIENAX, consistent across volume ranges.
Conclusions:
- NeuroQuant and MSmetrix provide cross-sectional WBV quantification with statistical agreement comparable to SIENAX.
- SIENAX is a well-validated tool for cross-sectional WBV assessment in MS clinical studies.
- These findings support the use of NeuroQuant and MSmetrix for WBV analysis in MS research.
